How to Write an Abstract
An abstract is a short, self-contained summary of your paper. To write one, finish the paper first. Then write one or two sentences each on the problem, the methods, the results and what they mean, and cut until you meet the word limit. Use real findings, not promises such as "results will be discussed". APA's Abstract and Keywords Guide sets a limit of 250 words, but your journal or instructor's limit comes first.
Check what you have been asked for
Abstracts vary, so check the rules before you draft.
| Situation | What to expect |
|---|---|
| APA paper | Up to 250 words in one paragraph, unless told otherwise |
| Student paper in APA style | APA says abstracts are not usually required, so ask your instructor |
| Journal article | Follow the journal's word limit and format, which may be structured |
| Conference | Follow the call for papers, which may be strict on length |
| Dissertation | Follow your institution's handbook |
The UNC Writing Center describes two broad types.
- Informative abstracts are the most common. They present the main arguments, key results and conclusions, so a reader gets the substance of the work.
- Descriptive abstracts say what the work contains, such as its purpose, methods and scope, but leave out results and conclusions. They are shorter.
If nobody specifies a type, write an informative abstract. Readers use it to decide whether to read on, and databases use its words for indexing.
What to include
APA lists the content usually needed, at about one or two sentences per topic. Here it is as a table, using one invented study of walking breaks and attention.
| Part | Question to answer | Example |
|---|---|---|
| Background | Why does this matter? | "Afternoon attention lapses are common in office work." |
| Problem or aim | What did you ask? | "We tested whether short walking breaks improve afternoon attention." |
| Hypothesis, if any | What did you predict? | "We expected walkers to make fewer errors." |
| Methods | What did you do, with whom? | "Sixty office workers were randomly assigned to walking or seated breaks for two weeks." |
| Results | What did you find? | "Walkers made fewer errors on an attention test." |
| Implications | So what? | "Short walking breaks may be a simple workplace option." |
Humanities abstracts often work differently. UNC notes that they tend to include the thesis, the background and the conclusion, rather than methods and results.
How to write it in six steps
Step 1: Finish the paper
You cannot summarize what you have not yet found. A draft abstract written early can help you focus, but rewrite it at the end.
Step 2: Reverse outline
UNC suggests summarizing the main idea of each section in a single sentence. Go through your paper and do this. Those sentences are your raw material.
Step 3: Draft without watching the word count
Write all the parts in the table above. Put the most important information first within each part.
Step 4: State your real results
This is where most abstracts are weakest. Compare:
- Weak: "The results of the study are presented and discussed."
- Better: "Walkers made fewer errors than seated participants on the attention test."
- Best, if the numbers are in your paper: "Walkers made fewer errors (M = 4.1, SD = 1.2) than seated participants (M = 6.3, SD = 1.5), p = .01."
Step 5: Cut
Remove background that any reader in the field would know. Remove phrases such as "this study aimed to investigate". Remove references and abbreviations you do not need. UNC advises revising repeatedly until you reach clarity and completeness in the fewest words.
Step 6: Check it against the paper
Read the abstract with your paper open. Check each number and claim. An abstract should not add anything that is not in the work.
A worked example
Here is a complete abstract for the invented study. The study and all its numbers are made up.
Unstructured version
Afternoon attention lapses are common in office work and can affect accuracy. This study tested whether short walking breaks improve afternoon attention more than seated breaks. We randomly assigned 60 office workers to take two 10-minute walking breaks or two 10-minute seated breaks each afternoon for two weeks. Attention was measured at the end of the workday with a computer-based error test. Walkers made fewer errors (M = 4.1, SD = 1.2) than seated participants (M = 6.3, SD = 1.5), p = .01. Self-reported fatigue did not differ between groups. Short walking breaks may offer a low-cost way to support afternoon attention at work. The small sample and short follow-up limit these conclusions, and larger trials over longer periods are needed.
That is about 120 words. Notice that it gives the design, sample, measure, numbers, a limit and an implication.
Structured version
Many health and science journals ask for labeled sections. The US National Library of Medicine explains structured abstracts and notes that labels such as these are common.
Introduction: Afternoon attention lapses are common in office work.
Methods: We randomly assigned 60 office workers to two daily 10-minute walking or seated breaks for two weeks, and measured attention with a computer-based error test.
Results: Walkers made fewer errors (M = 4.1, SD = 1.2) than seated participants (M = 6.3, SD = 1.5), p = .01. Fatigue did not differ.
Discussion: Short walking breaks may support afternoon attention. Larger, longer trials are needed.
Reporting guidelines for abstracts
Some study types have their own abstract checklists. PRISMA for Abstracts covers systematic reviews, and its checklist has 12 items, from title and objectives to funding and registration. CONSORT for abstracts covers randomized trials in journal and conference abstracts. If one applies, use it. It may be expected.
Formatting in APA 7
If you are using APA format, the guide gives these rules:
- Place the abstract on the second page of the paper.
- Write the label "Abstract" centered and in bold on the first line.
- Start the text one line below, as one double-spaced paragraph with no indent.
Keywords
Keywords are used for database indexing, and as search terms for readers. APA asks for three to five words, phrases or acronyms covering points such as the topic, population, method and application.
Format them one line below the abstract. Indent the label Keywords: in italics, then list them in lowercase, unless a word is a proper noun. Separate them with commas and leave off the final period.
Keywords: walking breaks, attention, office workers, randomized experiment
Choose terms a searcher would actually use. Do not just repeat your title.
Using an AI tool
If your rules allow it, CitedEvidence's abstract generator writes an abstract or summary from your own uploaded document. That helps if you are stuck on a first draft. Check your institution's or journal's AI rules first, and disclose use if required. Compare each number and claim with your paper. You are responsible for the final text.
Common mistakes
- Writing it first and never updating it. Your abstract then describes the paper you planned.
- Leaving out results. Readers come for your findings.
- Copying sentences from your introduction. Rewrite so that the abstract stands alone.
- Adding new information. If it is not in the paper, it does not belong here.
- Overstating. Do not turn "associated with" into "causes".
- Citing other works heavily. UNC advises against referring extensively to other work in an abstract.
- Defining terms at length. Use key terms and leave definitions to the paper.
- Missing the limit. Submission systems may enforce it. Count before you upload.
Before you submit, run the wider checks in our pre-submission checklist. If this is for a thesis, see our guide on improving your dissertation's structure and quality.
FAQ
How long should an abstract be?
Follow your journal, conference or instructor. APA's limit is 250 words unless you are told otherwise. UNC notes that descriptive abstracts usually run to 100 words or fewer, while informative abstracts usually run to no more than about 10% of the full work.
Should I write the abstract first or last?
Last. You can draft a rough version early to focus your thinking, but your final abstract must reflect the finished paper, including its actual results.
Can I include citations in an abstract?
Usually not. An abstract should stand alone and should not refer extensively to other works. If you must mention one, for example in a replication study, check the journal's rules.